Using personalized MSK models with MRI-based cartilage surfaces illustrated clearer differences in cumulative contact pressure between tibiofemoral compartments than generic models, though both approaches captured changes in knee contact mechanics due to gait biofeedback in ACLR subjects.
Key Findings
Results
Gait biofeedback targeting increased first peak vertical ground reaction force (vGRF) altered knee joint mechanics during walking in ACLR subjects.
Eight anterior cruciate ligament-reconstructed (ACLR) subjects were studied.
Biofeedback condition used visual feedback to increase the first peak vertical ground reaction force (vGRF) during walking.
The vGRF differed between habitual and biofeedback walking conditions.
Flexion angle, joint contact forces, and extensor moment also differed between walking conditions using either generic or personalized simulation approaches.
Results
The personalized musculoskeletal model showed clearer differences in cumulative contact pressure between tibiofemoral compartments compared to the generic model.
Personalized approach incorporated MRI-based cartilage articulating surfaces and ligament insertion points of the ACLR knees.
The personalized approach illustrated clearer differences in cumulative contact pressure between tibiofemoral compartments (p < 0.05, effect size ES > 1).
The generic model used a literature-based MSK model with a 12-degree-of-freedom knee.
The personalized model incorporated subject-specific geometry from magnetic resonance imaging (MRI).
Results
Both generic and personalized models detected differences in the excursion of center of contact pressure in the lateral compartment between walking conditions.
Differences were found in the excursion of the center of contact pressure in the lateral compartment of the femur (p < 0.05, ES > 0.5).
Differences were also found in the excursion of the center of contact pressure in the lateral compartment of the tibia (p < 0.05, ES > 0.7).
These differences between habitual and biofeedback walking were captured by both modeling approaches.
Analysis focused on the stance phase of walking.
Conclusions
A generic musculoskeletal model may suffice for limb-level kinetic and kinematic analysis in ACLR subjects.
Generic and personalized models both showed differences between habitual and biofeedback walking for vGRF, flexion angle, joint contact forces, and extensor moment.
The authors concluded that 'using a generic model may suffice for limb-level kinetic and kinematic analysis.'
However, the authors noted that 'validation is needed regarding joint site-specific conclusions.'
The study compared eight ACLR subjects across two walking conditions using both modeling approaches.
Background
ACLR patients exhibit less dynamic knee loading during walking than uninjured controls, which may be harmful to the joint.
Reduced dynamic knee loading following ACL reconstruction was identified as the clinical motivation for the study.
Increasing the first peak vGRF during walking by visual biofeedback was described as a strategy to promote more dynamic loading.
The authors framed reduced loading as potentially harmful to the joint, providing rationale for biofeedback intervention.
Prior musculoskeletal modeling work has simulated contact mechanics from this intervention, but impact of model personalization had not been examined.
Results
Model personalization affected the characterization of contact pressure distribution between tibiofemoral compartments more than limb-level kinematics and kinetics.
Both models agreed on limb-level outcomes such as vGRF, flexion angle, joint contact forces, and extensor moment.
The personalized model provided greater sensitivity for detecting compartment-specific pressure differences (ES > 1 vs. smaller or absent effects in generic model).
The study suggests joint site-specific conclusions require validation, implying generic models may not reliably characterize compartment-level contact mechanics.
Contact pressure-based variables from either modeling approach did capture changes in walking between habitual and biofeedback conditions.
What This Means
This research suggests that people who have had ACL reconstruction surgery tend to load their knee less during walking compared to uninjured individuals, and that this reduced loading could potentially be harmful to knee cartilage over time. Researchers used visual biofeedback — essentially showing participants real-time information about how hard they were pushing off the ground — to encourage more normal, dynamic knee loading. The study then compared two different computer modeling approaches to understand exactly how this biofeedback changed the pressures and forces inside the knee joint: one using a standard, 'off-the-shelf' model and one customized using MRI scans of each participant's own knee cartilage and ligament attachment points.
The study found that both the generic and personalized computer models successfully detected changes in overall knee mechanics (such as joint forces and leg movement angles) when participants used biofeedback compared to their normal walking. However, the personalized model — built from each person's own MRI data — was better at detecting differences in how pressure was distributed between the inner and outer compartments of the knee, showing larger and more statistically clear differences between compartments. Both models detected changes in where the center of pressure moved within the outer (lateral) compartment of the knee during biofeedback walking.
This research suggests that for measuring overall leg-level forces and movements, a standard computer model may be sufficient, but when researchers or clinicians want to understand what is happening at specific locations within the knee joint — such as whether one side of the joint is bearing more load than the other — a personalized model built from the individual's own MRI data may provide more accurate and meaningful information. The authors emphasize that further validation is needed before making specific clinical judgments based on joint site-specific model outputs, and that biofeedback-based gait training shows promise for modifying knee cartilage loading patterns in ACL-reconstructed individuals.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Paz A, Esrafilian A, Armitano-Lago C, Munsch A, Lisee C, Bjornsen E, et al.. (2026). Personalized musculoskeletal models show that gait biofeedback alters knee cartilage contact mechanics in ACL-reconstructed subjects.. Journal of biomechanics. https://doi.org/10.1016/j.jbiomech.2026.113376